Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

I. Introduction

Plastids are semi-autonomous plant organelles containing their own genome (plastid DNA; ptDNA). The compact 120–230-kb plastid genome encodes less than 100 proteins (Sugiura 1989; Raubeson and Jansen 2005); the majority of plastid functions is carried out by proteins encoded in ∼3,000 nuclear genes (Leister 2003). Plastid genes, transcription and translation have many conserved prokaryotic features (Barkan 2011).

Transformation of the plastid genome was first achieved in 1988 in the unicellular green alga Chlamydomonas reinhardtii (Boynton et al. 1988). Transformation of the plastid genome in tobacco (Nicotiana tabacum), a flowering plant species, followed in 1990 (Svab et al. 1990). Progress in Chlamydomonas plastome engineering has been the source of continued inspiration for researchers working with flowering plants. Shared features between the algal and flowering plant plastids are a polyploid genetic system, and reliance on nuclear genes for plastid function. However, the evolutionary distance is reflected in many mechanistic differences, and there is no expectation that protocols developed in either system would be interchangeable. The principal difference in the methodology can be traced back to engineering of the plastid genome of algal cells in photoautotrophic cultures and manipulation of the plastid genome in higher plants in heterotrophically grown tissue culture cells.

Since 1990 plastid transformation has been implemented in numerous flowering plant species. This review will focus on the methods for engineering the plastid genome of flowering plants and gives an overview of the progress made in implementing plastid transformation in different taxonomic groups. For information on the applications of plastid transformation in basic science and biotechnology, the reader is referred to recent reviews (Daniell et al. 2009; Cardi et al. 2010; Day and Goldschmidt-Clermont 2011; Maliga and Bock 2011; Whitney et al. 2011).

II. Methods for DNA Introduction

There are two practical methods of DNA introduction into plastids: biolistic DNA delivery and polyethylene glycol (PEG)-mediated DNA uptake.

A. Biolistic DNA Delivery

Protocols for biolistic delivery of RNA and DNA into living cells were developed by John Sanford’s laboratory. In the first experiments, delivery of tobacco mosaic virus RNA was confirmed by formation of viral inclusion bodies in onion cells (Klein et al. 1987) and transient expression of introduced nuclear reporter genes was confirmed by measuring CAT and GUS reporter enzymes in bombarded onion and maize tissue (Klein et al. 1987, 1988a). Stable genetic transformation of the tobacco nucleus (Klein et al. 1988b), yeast mitochondria (Johnston et al. 1988) and the chloroplasts in Chlamydomonas (Boynton et al. 1988; Blowers et al. 1989) and higher plants (Svab et al. 1990) followed in rapid succession. Early protocols for biolistic DNA delivery involved precipitation of the transforming DNA with CaCl2 and spermidine free base on the surface of microscopic (0.6–1.0 μm) tungsten or gold particles, and accelerating the particles using a gunpowder-charge driven device to speeds that enable penetration of multiple cell layers. Acceleration of particles was carried out in vacuum in the PDS-1000 gun and solid support of the bombarded cells was provided in the form of a filter paper facilitating particle penetration. All these important elements for success were identified early on (Klein et al. 1987, 1988a). A cleaner, more efficient device is PDS-1000/He in which helium replaces the role of the gunpowder charge (Ye et al. 1990). A useful recent addition to the PDS-1000/He device is the hepta adaptor enabling simultaneous bombardment with seven macrocarriers.

An alternative particle gun design is the Particle Inflow Gun (PIG) that also uses pressurized helium in combination with a partial vacuum to accelerate DNA-coated tungsten or gold particles (Finer et al. 1992). The particles in the PIG are accelerated directly in a helium stream rather than being supported by a macrocarrier, as in the PDS1000/He gun. Because the PIG is not available commercially, it is relatively rarely used. However, it appears to be as efficient as the PDS1000/He gun for plastid transformation (Dufourmantel et al. 2004, 2007).

The targets for plastid transformation by biolistic DNA delivery most often are plastids in leaves (Svab et al. 1990; Svab and Maliga 1993) or less frequently in tissue culture cells (Langbecker et al. 2004). Osmotic stabilizers in some instances are used to protect tissue culture cells during bombardment, although the efficiency of protection has not been rigorously proven (Langbecker et al. 2004).

Historically, biolistic DNA delivery to plastids was optimized using transient expression of GUS and CAT reporter enzymes expressed from plastid signals (Daniell et al. 1990; Ye et al. 1990); for review see (Sanford et al. 1993). Only a small fraction of overall activity detected in these experiments is likely to derive from plastids because genes in plastid cassettes are also expressed in the nucleus (Cornelissen and Vandewiele 1989). The nucleus is transformed 20–40-times more efficiently than plastids (Langbecker et al. 2004) and initial plastid expression from a few transformed ptDNA copies is only a fraction of protein levels measured at the homoplastomic state. Therefore, these experiments likely determined conditions for DNA delivery to the plant nucleus rather than to plastids. Protocols detecting DNA delivery to the nucleocytosolic compartment are still useful to identify conditions for plastid transformation, because delivery of DNA into the cell is sufficient to obtain plastid transformation (see PEG-mediated plastid transformation below). Only one systematic study of biolistic DNA delivery was carried out that measured the success of DNA delivery by the number of transplastomic clones (Langbecker et al. 2004). The number of transplastomic clones obtained with 0.6 or 1.0 μm particles in tissue culture cells and in leaves was comparable, ∼1 per bombarded sample. However, plastid transformation in tobacco tissue culture cells with the smaller 0.4 μm particles was 3–4-times more efficient that with the standard 0.6–1.0 μm particles, yielding ∼4 transplastomic clones per bombarded sample. Detailed protocols are available for biolistic transformation of tobacco leaf cells (Bock 2001; Lutz et al. 2006b; Lutz and Maliga 2007a; Maliga and Svab 2011) and tissue culture cells (Langbecker et al. 2004).

B. Polyethylene Glycol Treatment

Plastid transformation by polyethylene glycol (PEG) treatment of protoplasts utilizes the empiric DNA uptake process developed for nuclear gene transformation (Paszkowski et al. 1984). PEG treatment was first used to demonstrate transient expression of the introduced GUS reporter gene in isolated tobacco chloroplasts (Sporlein et al. 1991), followed by stable genetic transformation of the plastid genome in Nicotiana tabacum (Golds et al. 1993) and Nicotiana plumbaginifolia (O’Neill et al. 1993). A detailed protocol for PEG-mediated transformation of plastids in tobacco protoplasts is available (Koop et al. 1996).

Because of its ease of application, biolistic DNA delivery is by far the most frequently used method for plastid transformation. Protoplast isolation, PEG treatment and plant regeneration from protoplasts require more training and are more laborious and time-consuming. However, plastid transformation by PEG treatment is in the public domain and does not require expensive equipment, thus it may be preferable to biolistic DNA delivery in some applications (Dix and Kavanagh 1995).

III. Marker Genes

The challenge of plastid transformation has been to uniformly alter the hundreds to thousands of plastid genome copies localized in ten to hundreds of organelles in a plant cell. DNA delivery produces only a few transformed ptDNA copies, which are then selectively amplified while the cells are grown in tissue culture. Selection for transformed plastid genomes is essential to recover genetically uniform transplastomic plants. Tobacco shoots regenerated from a bombarded leaf are always chimeric. Two cycles of plant regeneration on a selective medium, coupled with probing total cellular DNA for the uniformity of ptDNA, is typically sufficient to obtain genetically stable plants. Repeated cycles of plant regeneration are necessary, because cells in different developmental layers in a shoot apex may differ in their segregation patterns of the two plastid genome types. Regeneration of a new shoot apex from a small group of cells on a selective medium is used to obtain genetically uniform, homoplastomic plants (Lutz and Maliga 2008). Alternatively, visual-selective markers may track progress toward the homoplastomic state (Tungsuchat-Huang et al. 2011). Below is a review of the selectable marker genes that are available for the construction of transplastomic clones.

A. Primary Positive Selection

Detoxifying enzymes that enable the growth of cells on a normally toxic medium provide selective advantage to plastids so that they gradually outnumber non-transformed plastids in cells grown in culture. If the cellular target of antibiotic action is known, genes encoding insensitive forms of the cellular target may also be used as selective markers. The selective plastid markers fall in two classes: primary selective markers that confer a selective advantage early on, when only a few ptDNA copies are amplified; and secondary selective markers that confer protection only when a significant portion of ptDNA copies already carry the marker (see below).

Most primary selective agents are selective inhibitors of plastid protein synthesis on the prokaryotic type (70S) ribosomes, which do not affect mRNA translation on the eukaryotic 80S ribosomes in the cytoplasm. The group of antibiotics that can be used as a primary selective agent includes spectinomycin, streptomycin, kanamycin and chloramphenicol. These antibiotics inhibit greening, cell division and shoot formation in culture on a shoot regeneration medium. Transplastomic clones can be identified by the absence of phenotypes associated with antibiotic treatment of wild-type cells, that is, they show greening, faster proliferation and shoot formation on an antibiotic-containing plant regeneration medium. The first transplastomic clones were obtained by spectinomycin selection for mutant forms of the 16S rRNA, which do not bind the antibiotic (Svab et al. 1990; Staub and Maliga 1992, 1993). The mutant rrn16 genes in the plastid transformation vectors were soon replaced with the more efficient aadA gene encoding aminoglycoside 3″-adenylyltransferase or AAD (Svab and Maliga 1993). AAD inactivates both spectinomycin and streptomycin. Resistance to both antibiotics is exploited to distinguish relatively frequent spontaneous spectinomycin resistant mutants from transplastomic clones, because only transplastomic clones, but not plastid rRNA mutants, are resistant to both antibiotics.

Kanamycin resistance has also been suitable to recover transplastomic clones. The first plastid-engineered kanamycin resistance (neo) genes were relatively inefficient (Carrer et al. 1993), but increasing expression of the encoded enzyme neomycin phosphotransferase II (NPTII) yielded marker gene variants that are as efficient as aadA, yielding about one transplastomic clone per bombarded sample (Lutz et al. 2004). Kanamycin resistant clones were also recovered by selection for the aph(3′)IIa gene (Huang et al. 2002).

There are two recent additions to the primary selective plastid markers, both of which were tested in tobacco. One of the new selective agents is chloramphenicol, inhibiting translation on plastid ribosomes as do spectinomycin and kanamycin. Chloramphenicol resistance appears to be less robust than spectinomycin or kanamycin resistance, because selection in tobacco should be carried out in low light and the color change is more subtle (Li et al. 2011). A distinct advantage of the marker is the absence of spontaneous chloramphenicol resistance mutants. The second marker system explored selection for the feedback-insensitive anthranilate synthase (AS) alpha-subunit gene of tobacco (ASA2) that ­confers resistance to the indole analogue 4-­methylindole (4MI) or the tryptophan analogue 7-methyl-DL-tryptophan (7MT) (Barone et al. 2009). Testing of the new markers in additional plant species will be necessary to fully assess their utility.

Selection for betaine aldehyde (BA) resistance after transformation with a vector carrying a spinach betaine aldehyde dehydrogenase (badh) gene was reported to be efficient for the recovery of transplastomic clones (Daniell et al. 2001; Verma and Daniell 2007). The betaine aldehyde dehydrogenase enzyme (BADH) converts toxic BA to betaine, an osmoprotectant accumulating in some plants in dry or saline environments. Attempts to duplicate the selection protocol in other laboratories were unsuccessful, as discussed in a recent review (Maliga 2004). Because no plants were described in the literature that carry badh as the only selective marker (badh was always combined with aadA), for the time being, badh should be considered a putative marker only.

B. Secondary Positive Selection

Protection conferred to plant cells by secondary selective markers is dose dependent. These markers are not suitable to enrich for transplastomic plastids when only a few ptDNA copies are transformed, but will confer a selective advantage when many or most genome copies carry the marker gene. Examples for secondary selective marker genes are those that confer resistance to the herbicides phosphinothricin (PPT; (Lutz et al. 2001; Ye et al. 2003)), glyphosate (Ye et al. 2003), sulfonylurea, pyrimidinylcarboxylate (Shimizu et al. 2008) and diketonitrile (Dufourmantel et al. 2007). Low level expression of the protective enzyme from the few initially transformed ptDNA copies, as opposed to full expression from a nuclear transgene may explain why these markers are suitable to directly recover nuclear transformants, but require enrichment to recover transplastomic clones. Subcellular localization of the protective enzymatic activity may also be a contributing factor.

Actinonin is a selective and potent inhibitor of plant peptide deformylases (Fernandez-San Millan et al. 2011). Expression of the Arabidopsis thaliana peptide deformylase PDF1B (linked to spectinomycin resistance) in tobacco chloroplasts conferred actinonin resistance to the transformed plants. However, when the combination of the PDF1B gene and actinonin was used as the primary selective marker system for chloroplast transformation, all developed shoots were escapes. Therefore, the use of this system would be limited to the role of a secondary selective marker (Fernandez-San Millan et al. 2011).

C. Negative Selection

Negative selection is also available in plastids. It selects for the loss of a conditionally toxic gene. Negative selection in plastids is based on the expression of the cytosine deaminase enzyme making the cells sensitive to 5-fluorocytosine. The loss of the bacterial codA gene (encoding cytosine deaminase) could be detected by cellular proliferation on 5-fluorocytosine-containing medium (Serino and Maliga 1997; Corneille et al. 2001).

D. Visual Plastid Marker Systems

Because the plants that are expressing selectable marker gene have no visual phenotype, the ­uniform transformation of plastid genomes (= homoplastomic state) can be verified only by DNA gel blot analyses and the absence of segregation in the seed progeny. Since deletion of most plastid genes causes a dramatic change in leaf color, changes in chlorophyll content have been utilized as a marker system to facilitate rapid identification of plastid genotypes. The Koop laboratory (Klaus et al. 2003) developed a system for the rapid identification of transplastomic sectors using pigment-deficient tobacco knockout plants as recipients. In the knockout plants, the first plastid marker (aadA, encoding spectinomycin resistance) replaces a plastid gene that causes chlorophyll deficiency. The second transformation vector carries the photosynthetic gene to restore green pigmentation linked to a second marker (aphA-6, encoding kanamycin resistance). Homoplastomic sectors and plants can be readily identified by the restoration of green pigmentation among plants selected for kanamycin resistance.

Variants of this protocol have been developed that require only one selectable marker and are directed towards manipulation of rbcL, the plastid-encoded Rubisco large subunit gene in tobacco. In one approach (Kode et al. 2006), deletion of the plastid rbcL gene was obtained by homology-based deletion using a two-step protocol. First, selection for spectinomycin resistance (aadA) was used to duplicate the rbcL flanking sequence. Subsequently, deletion of rbcL and the linked aadA by a (spontaneously occurring) homologous recombination event was recognized in the seed progeny by appearance of the pigment-deficient phenotype. The rbcL deletion line could subsequently be transformed with a functional rbcL allele linked to aadA. The homoplastomic sectors (plants) could be readily identified by their green pigmentation. In a variant approach (Whitney and Sharwood 2008), the tobacco rbcL gene was replaced with a heterologous rbcL sequence using aadA as a selective marker. The aadA gene was subsequently removed by the Cre site-specific recombinase, so the master line was ready to be transformed with rbcL variants using aadA as a selective marker.

The visual marker system discussed above relies on pigment deficiency caused by a missing or defective plastid gene. Our novel visual marker system relies on interference of a plastid transgene with the expression of the clpP plastid gene. The transgene acts as a “poison pill” because it contains a clpP segment that interferes with the maturation of the native clpP mRNA (Kuroda and Maliga 2002). So far, two variants of the visual marker have been tested: the aurea bar (bar au) (Kittiwongwattana et al. 2007; Lutz and Maliga 2008) and aadA au (Tungsuchat-Huang et al. 2011) transgenes conferring a golden leaf phenotype to plants. Because the bar au gene is not a primary selectable marker, its deployment requires two genes: the aurea bar (bar au) gene that confers a golden leaf phenotype and a spectinomycin resistance (aadA) gene that is necessary for the introduction of the bar au gene in the plastid genome. The aadA au transgene fulfills both functions: it is a conventional selectable aadA gene in culture, and allows detection of transplastomic sectors in the greenhouse by leaf color. Because the aurea plants are viable, the aurea plastid genes are useful to query rare events in large populations (Tungsuchat-Huang et al. 2010).

E. Reporter Genes

The E. coli β-glucuronidase (GUS) reporter enzyme facilitates the monitoring of gene expression, but does not confer a selective advantage or disadvantage to plastids. GUS enzymatic activity expressed in chloroplasts has been measured using fluorogenic assays (Staub and Maliga 1993, 1994; Eibl et al. 1999; Zou et al. 2003) and visualized by histochemical staining (Staub and Maliga 1993; Iamtham and Day 2000; Zubko et al. 2004; Sheppard et al. 2008).

The Aequorea victoria green fluorescent protein (GFP) is a visual marker, allowing direct imaging of the fluorescent gene product in living cells. Its chromophore forms autocatalytically in the presence of oxygen and fluoresces green when absorbing blue or UV light. GFP has been used to detect transient gene expression (Hibberd et al. 1998) and stable transformation events (Sidorov et al. 1999; Shiina et al. 2000; Reed et al. 2001) in chloroplasts. GFP-expressing chloroplasts in tissue grafts facilitated demonstration of the transfer of genetic material between cells (Stegemann and Bock 2009).

GFP was fused with AAD, the aadA gene product that confers spectinomycin resistance, to be used as a bifunctional visual and selective (spectinomycin resistance) marker gene (Khan and Maliga 1999). Transformation vectors carrying the aadA-gfp marker gene were used to recover stable transplastomic clones in N. tabacum (Khan and Maliga 1999), N. sylvestris (Maliga and Svab 2011) and Lesquerella fendleri (Skarjinskaia et al. 2003).

Luciferases are enzymes that emit light in the presence of oxygen and a substrate (luciferin) and which have been used for real-time, low-light imaging of gene expression in cell cultures, individual cells, whole organisms, and transgenic organisms. Luciferases have served as reporters in a number of promoter search and targeted gene expression experiments over the last two decades (Greer and Szalay 2002). Until now, expression of various luciferases in plants has required exogenous application of luciferins – frequently toxic and high-cost ­compounds – to achieve only temporary and relatively low light emission levels from live plant tissues. Evolutionary conservation of the prokaryotic gene expression machinery enabled expression of the six genes of the lux operon in chloroplasts yielding plants that are capable of autonomous light emission (Krichevsky et al. 2010). This system now can be modified for gene expression studies and for genetic screens.

IV. Vectors

Plastid transformation vectors consist of a vector backbone for cloning and propagation in E. coli, a plastid targeting region with a selectable plastid marker to facilitate integration of the gene-of-interest into the plastid genome, and optional sequences to facilitate marker gene excision. The vector backbones are pUC or pBluescript plasmid derivatives carrying a ColE1 replication origin that ensures plasmid replication in E. coli but not in plastids. Because the ColE1 replication origin does not function in plastids, the plastid marker is expressed in the plant cell only if it integrates into the plastid genome. The pUC and pBluescript vectors encode ampicillin resistance as the selectable marker in E. coli, which is not a suitable selectable marker in plastids. Spectinomycin, kanamycin or chloramphenicol resistance genes engineered for expression in plastids are also selectable in E. coli, therefore dual selection for the bacterial ampicillin resistance and the plastid marker ensures maintenance of intact (deletion-free) copies of plastid vectors.

The plastid-targeting region is a ∼0.5–2.0-kb ptDNA fragment flanking the marker gene (and gene of interest) to facilitate integration of the marker gene (and the gene-of-interest) into the ptDNA by two homologous recombination events. The vector design is dependent on the desired ptDNA manipulation that can be insertion of foreign genes, replacement of native plastid genes with mutant forms, gene deletion or cotransformation.

A. Insertion Vectors

Expression of transgenes requires plastid insertion vectors that enable convenient DNA manipulation in E. coli and targeted insertion of the gene-of-interest into the plastid genome. Because the insertion vectors are repeatedly used for the insertion of different genes, significant effort has been invested to characterize the insertion site in the plastid genome and endow the vectors with convenient features. Characterization of the insertion site includes, for example, ensuring that there is no interference with the expression of adjacent plastid genes and identification of read-through transcripts that may enhance or reduce transgene expression. Vector convenience features are, for example, convenient restriction sites for cloning, alternative selection markers, and sequences to facilitate post-transformation removal of marker genes. Because vector development requires a significant effort, only a few vectors are used routinely. The pRB94/95 vectors (Ruf et al. 2001) and our pSS24/25 vectors (Sinagawa-Garcia et al. 2009) target transgenes in the single-copy region of the plastid genome, whereas our pPRV vector series (Zoubenko et al. 1994; Lutz et al. 2007) and the pSBL-CTV2 vectors (Daniell et al. 1998) target insertions in the repeated region of the plastid genome. Insertion of transgenes in the repeated region yields ptDNA with two transgene copies per genome.

When choosing plastid-targeting sequences for vector construction, DNA sequence variation within species and between species is a concern. Ideally, vectors should contain sequences identical to the target ptDNA for optimal recombination. Targeting regions with point mutations in synthetic DNA behave as homologous sequences; the recombination sites are at either ends of the targeting region (Sinagawa-Garcia et al. 2009). Some degree of sequence variation is tolerated as long as sufficiently extensive regions of homology are present. In a now classic study, transformation of N. tabacum plastids with Solanum nigrum vectors has shown that transformation with 97.6% similar (homeologous) sequences (sequence divergence 2.4%) is as efficient as with identical sequences (Kavanagh et al. 1999). Vectors with N. tabacum targeting sequences are used to transform plastids in potato (Sidorov et al. 1999), tomato (Ruf et al. 2001), petunia (Zubko et al. 2004) and N. sylvestris (Maliga and Svab 2011). The plastid genomes of the amphiploid species Nicotiana tabacum and its maternal progenitor N. sylvestris differ only by seven sites: three in introns, two in spacer regions and two in coding regions (Yukawa et al. 2006). None of the known differences are within the plastid targeting regions of our standard pPRV or pSS24/25 vectors and, even if they were, the point mutations and insertions/deletions (affecting one or two nucleotides) would not significantly affect transformation efficiency. However, replacement of tobacco-specific vectors (sequence divergence 4.6%) with potato-specific vectors increased potato plastid transformation efficiency 10-fold (Valkov et al. 2011). Thus, construction of species-specific, or even line-specific, vectors is advisable, if there is significant intraspecific variation in the ptDNA. Sequencing the plastid genomes of two tomato cultivars (IPA-6 and Ailsa Craig) revealed that they are identical to the nucleotide (Kahlau et al. 2006); thus, one vector for tomato should be sufficient. However, significant sequence variation in the ptDNAs of rice subspecies (Tang et al. 2004) may justify construction of multiple plastid transformation vectors for rice.

There is only limited information on the importance of choosing homologous expression signals for transgene expression. In most plastid transformation vectors the marker genes are driven by the “heterologous” tobacco rrn operon PEP promoter. Because the rrn PEP promoter elements are conserved between dicots and monocots (with the only know exception being spinach; (Sriraman et al. 1998; Suzuki et al. 2003)), this promoter is not really heterologous. However, the efficiency of expressing recombinant proteins from the psbA promoter appears species specific (Ruhlman et al. 2010). Systematic testing of the utility of expression signals in heterologous systems will be an important area for future research.

The general insertion vectors have only a marker gene and a linked multicloning site. Specialized vectors, in addition, have a gene of interest on which one element, for example the promoter, can be readily exchanged to create a series of constructs. Such specialized vectors are the vectors developed to study plastid RNA editing. Three approaches were used. Conceptually the simplest design was construction of minigenes that were obtained by inserting in a plastid expression cassette a DNA fragment that contains (an) editing site(s) (Reed and Hanson 1997). The second approach, translational fusion with a reporter gene was used to study the psbL and ndhD editing events that create an AUG translation initiation codon by editing of an ACG codon at the mRNA level (Chaudhuri and Maliga 1996). The third approach was incorporation of editing segments in the 3′UTR of the aadA marker gene where the editing status of the segment does not affect expression of the marker gene (Bock et al. 1996). For a review of plastid editing vectors, see (Lutz and Maliga 2007a).

B. Replacement Vectors

Replacement vectors are variants of insertion vectors, when the sequence to be inserted is already present in the ptDNA and the intent is to replace the native sequence with a variant gene (mutant allele) incorporated in the vector targeting region. Replacement vectors are individually tailored to engineer specific genes. Replacement vectors have been developed for engineering rbcL, the gene encoding the large subunit of the Rubisco enzyme. Significant similarity between the native sequence and the variant, such as the tobacco and sunflower rbcL genes allowed undesirable recombination within the rbcL gene (Kanevski et al. 1999). To avoid this, the target gene sequence was either deleted (Klaus et al. 2003) or replaced with a dissimilar sequence (Whitney and Sharwood 2008), and the knockout/engineered plant is then used a master recipient for gene replacement. Efficient recovery of transplastomic clones was facilitated by restoration of green pigmentation, as discussed in Sect. III.D.

C. Deletion Vectors

Deletion vectors are designed to create knockout lines lacking specific plastid genes by replacing the target gene with a selectable marker gene by homologous recombination via the flanking ptDNA sequences. Knockout lines could be obtained for most plastid genes. For example, deletion of the plastid rbcL or rpoB genes makes the plants pigment deficient, but the knockout plants can be maintained on sucrose-containing medium or by grafting onto wild-type plants. In some instances, for example in the case of the plastid ndh genes, the knockout phenotype does not significantly interfere with photosynthesis and viability, while in other cases, for example clpP1, the plastid genes are essential for viability even on sucrose-containing medium. For reviews see (Bock 2001; Maliga 2004) and Chap. 18 in this volume.

D. Cotransformation

Cotransformation is a process when transformation is carried out with two (or more) vectors, targeting multiple regions of the plastid genome. At least one of the vectors carries a selectable marker gene so that transplastomic clones can be recovered by selection. Because bombardment is carried out with mixed plasmids and integration of both plasmids is efficient, ∼20% of the clones selected by the antibiotic resistance encoded in one vector will carry integrated copies of the second vector lacking a selectable marker (Carrer and Maliga 1995). Cotransformation has been exploited to tag an unlinked ndh gene (Rumeau et al. 2005) and to obtain marker-free herbicide resistance plants (Sect. V.D, Ye et al. 2003).

V. Marker Excision

The marker genes are essential for the selective enrichment of rare transformed ptDNA copies. However, when uniform transformation of ptDNA copies is achieved, the marker gene is no longer necessary to maintain the transplastomic state. Reasons for posttransformation removal of marker genes are: the shortage of primary selectable markers (spectinomycin selection for aadA is by far the most convenient), high-level expression of the marker genes imposing a metabolic burden on the plant, and consumer acceptance. There are four principal protocols for marker excision, each of which requires a special vector design discussed below. For reviews, see (Lutz and Maliga 2007b; Day and Goldschmidt-Clermont 2011).

A. Repeat-Mediated Excision

Repeat-mediated marker excision, developed in Anil Day’s laboratory, requires flanking the sequence targeted for deletion by a duplicated segment of at least a few hundred base pairs. The duplicated structure is unstable, and homologous recombination will eventually result in deletion of the sequence between the repeats. The advantage of homology-based marker excision is that it is seamless, leaving behind no extraneous sequence. However, repeat-mediated marker excision is difficult to control, because deletion may take place in E. coli during cloning or during transformation before reaching the homoplastomic state (Iamtham and Day 2000; Day and Goldschmidt-Clermont 2011). Homology-based marker excision has been used in soybean to obtain marker-free herbicide-resistant plants (Dufourmantel et al. 2007).

B. Excision by Phage Recombinases

Marker excision by phage site-specific recombinases is a two-step process: first, transplastomic plants are obtained in the absence of recombinases and, when marker excision is desired, plastid-targeted recombinases are expressed in the cells (Lutz and Maliga 2007b). To set up the lines for marker excision, the P1 phage loxP site (Corneille et al. 2001; Hajdukiewicz et al. 2001) or the phiC31 phage attP/attB sites (Kittiwongwattana et al. 2007) flank the marker genes in the plastid transformation vectors. The plastid genomes carrying target site-flanked marker genes are stable in the absence of recombinases (Tungsuchat-Huang et al. 2010). However, excision of the marker genes is very efficient when the gene of the plastid-targeted recombinase is introduced into the nuclear genome by transformation or crossing (Corneille et al. 2001; Hajdukiewicz et al. 2001; Kittiwongwattana et al. 2007), or transiently from Agrobacterium T-DNA (Lutz et al. 2006a). When using phage site-specific recombinases, a copy of the recombinant target site is left behind in the plastid genome.

C. Transient Cointegration

The third approach is the so-called transient cointegration protocol, in which the marker gene is outside the plastid targeting region of the transformation vector (Klaus et al. 2004). Placing the marker gene outside the targeting region enables selection for a cointegrate structure that forms by recombination between the ptDNA and the transformation vector via only one of the plastid targeting regions. As the result, the entire vector is incorporated in the ptDNA. When selection for the antibiotic resistance marker is stopped, recombination via the second targeting region can take place and the marker gene is excised. This marker excision system is also seamless, and antibiotic selection provides a degree of control.

D. Cotransformation and Segregation

Marker-free herbicide resistance plants have been obtained after transformation with mixed plasmids and a consecutive two-step selection process (Ye et al. 2003). The transformed plastids were first selected on spectinomycin-containing medium to identify clones, which were grown from cells bombarded with mixed plastids. A significant fraction of plastid genome copies in these cells carried integrated herbicide resistance genes targeted to a second integration site. Glyphosate or phosphinothricin are not suitable for the recovery of transplastomic clones when present in only a few copies in a cell, as discussed in Sect. III.B. However, spectinomycin resistance enabled propagation of integrated herbicide-resistance genes so that they could be directly selected for during a second cycle of plant regeneration. Some of the ptDNA copies carrying integrated herbicide resistance genes do not have integrated copies of the spectinomycin resistance gene, thus enabling ­segregation of spectinomycin marker-free plants (Ye et al. 2003).

VI. Flowering Plant Species with Systems for Plastid Transformation

Identification of transplastomic tobacco lines is based on two general criteria: greening of transplastomic cells (chlorophyll accumulation) on the selective medium that normally inhibits growth and chlorophyll accumulation, and capacity for regeneration from cultured cells so that homoplastomic cells can be obtained during repeated cycles of plant regeneration. The key to extending plastid transformation to new species has been combining a species-specific regeneration protocol with antibiotic treatment that blocks greening and tissue proliferation. Below is a brief review of the state of the art of plastid transformation in the different taxonomic groups. Highlighted in the crop species section will be (1) the laboratories making significant contributions to technology development, (2) the choice of methods for DNA introduction, (3) the marker genes used for selection, (4) the cultivars or accession in which the methods have been tested, (5) the salient features of the system and (6) its main uses.

A. Tobacco: Nicotiana tabacum and Other Species in the Genus Nicotiana

N. tabacum cv. Petit Havana was the first tobacco cultivar in which we reported plastid transformation with a mutant rrn16 gene in 1990 (Svab et al. 1990). The recessive rrn16 gene was soon replaced with the dominant aadA gene that is more efficient yielding about one transplastomic clone per bombarded sample (Svab and Maliga 1993). To date, virtually all tools and protocols for plastid transformation have been developed using this cultivar (for details, see sections above). The most commonly used protocols employ shoot regeneration from bombarded leaf tissue (Lutz et al. 2006b; Lutz and Maliga 2007a), although a protocol for transforming proplastids in tissue culture cells was also described (Langbecker et al. 2004). Plastid transformation in other Nicotiana species with a similar tissue culture response could be readily duplicated using N. tabacum cv. Petit Havana protocols, including Nicotiana plumbaginifolia (O’Neill et al. 1993), Nicotiana benthamiana (Davarpanah et al. 2009) and Nicotiana sylvestris TW137 (Maliga and Svab 2011). The cv. Petit Havana plants are relatively small and flower early. To obtain plants with a larger biomass, plastid transformation has been extended to additional tobacco cultivars, including Wisconsin 38 (Iamtham and Day 2000), Xanthi, Burley (Lee et al. 2003), Samsun, K327 (22X-1; (Yu et al. 2007)) and Maryland Mammoth (McCabe et al. 2008). N. tabacum is the model species of plastome engineering and is widely used in basic science studies and for biotechnological applications (Daniell et al. 2009; Cardi et al. 2010; Day and Goldschmidt-Clermont 2011; Maliga and Bock 2011; Whitney et al. 2011).

B. Potato: Solanum tuberosum

Plastid transformation in potato was reported by the Monsanto group (Sidorov et al. 1999) in FL1607, a highly regenerable, non-commercial potato line. Transformation was carried out with tobacco-specific vectors, which carried tobacco ptDNA fragments to target insertions into the potato ptDNA. The vectors carried aadA as a selectable marker and shoot regeneration was carried out in the presence of spectinomycin (300 mg/L). The yield of transplastomic clones was lower than in tobacco, one transplastomic clone in 15–30 bombarded leaf samples. Comparably low plastid transformation efficiency was obtained with the aadA marker gene, spectinomycin selection (300 mg/L) and tobacco-specific vectors in Solanum tuberosum cv. Desiree, a commercial cultivar (Nguyen et al. 2005). Transplastomic clones in FL1607 were recovered in a single-step regeneration protocol as in tobacco. In cv. Desiree, a two-step procedure was adopted: selection was first carried out on a callus-induction medium, then on shoot-induction medium. A dramatic, ∼10-fold increase in transformation efficiency was obtained when the tobacco-specific targeting sequences were replaced with potato-specific targeting sequences in cv. Desiree, using an improved two-step procedure yielding about one transplastomic clone per bombarded sample (Valkov et al. 2011). Leaf bombardment was carried out on a medium containing 0.1 M sorbitol and 0.1 M mannitol as osmoticum. Because the transplastomic clones were grown for a long time (3–4 months) as callus before plant regeneration, almost all (92%) of the regenerated plants were homoplastomic. GFP in transplastomic leaves accumulated up to 3–5% of total soluble protein as compared to 0.02–0.05% in tubers (Sidorov et al. 1999; Valkov et al. 2011) indicating that optimization of protein expression is required if expression of recombinant proteins in potato tuber amyloplasts is the goal.

C. Tomato: Solanum lycopersicum

Plastid transformation in tomato has been developed in Ralph Bock’s laboratory using biolistic DNA delivery, tobacco-specific vectors carrying the aadA marker gene and spectinomycin selection (500 mg/L; (Ruf et al. 2001)). Transformation has been carried out in two South American varieties: Santa Clara and IPA-6 (Wurbs et al. 2007; Zhou et al. 2008; Apel and Bock 2009). Plastid transformation in tomato has also been obtained by PEG-treatment of protoplasts, using tobacco (N. tabacum) or Solanum nigrum-specific vectors carrying binding-type spectinomycin and streptomycin resistance markers in the rrn16 genes and selection for spectinomycin resistance (300 mg/L). Transformation was carried out in the tomato processing cultivar T1783 (Nugent et al. 2005). Although plastid transformation in tomato has been significantly improved over time (Wurbs et al. 2007; Zhou et al. 2008), initial construct optimization in the well-established tobacco system is advisable.

Applications of tomato plastid transformation include engineering the carotenoid metabolic pathway and expression of antigens for subunit vaccines (Wurbs et al. 2007; Zhou et al. 2008; Apel and Bock 2009). Some of the recombinant proteins (p24-Nef) accumulated to up to 40% of the total soluble cellular protein in tomato leaves, but no significant protein accumulation was detected in ripe tomato fruits suggesting that protein expression in chromoplasts will require a specialized expression system (Zhou et al. 2008). The presumably relatively low enzyme levels were sufficient for successful metabolic pathway engineering (Wurbs et al. 2007; Apel and Bock 2009).

D. Petunia: Petunia hybrida

Plastid transformation in petunia has been reported from Anil Day’s laboratory (Zubko et al. 2004). Tobacco-specific transformation vectors carrying an aadA gene were introduced into petunia leaves by the biolistic process, and transplastomic shoots were regenerated on a medium containing spectinomycin (200 mg/L) and streptomycin (200 mg/L). Transformation was carried out in the Pink Waive commercial cultivar. Petunia hybrida is a diploid species that is suitable to study the biology of flowering plants using transgenic approaches (Gerats and Vandenbussche 2005; Gillman et al. 2009). Thus, applications of plastid transformation in Petunia are expected to follow.

E. Eggplant: Solanum melongena

Plastid transformation in eggplant was developed in K.C. Bansal’s laboratory (Singh et al. 2010). Tobacco-specific vectors carrying the aadA marker gene were introduced into green stem segments by the biolistic process and transplastomic shoots were regenerated on spectinomycin-containing medium (300 mg/L) using a one-step protocol. Initial selection on spectinomycin was followed up by selection for spectinomycin and streptomycin (300 mg/L each). Plastid transformation was essentially carried out as in tobacco, except that the transforming DNA was introduced into green stem segments instead of leaves.

F. Soybean: Gycine max

Soybean was the first major agronomic crop in which plastid transformation was implemented by a group of researchers at Bayer Crop Science (Dufourmantel et al. 2004). Plastid transformation was achieved by biolistic delivery of soybean-specific vectors carrying an aadA gene and the transplastomic clones were identified by their green color on spectinomycin medium in cv. Jack. The Bayer group used the particle inflow gun (PIG), rather than the DuPont biolistic gun. Noteworthy about the soybean system is that the transforming DNA was introduced into green embryogenic calli. The green embryogenic callus bleached in the presence of 200 or 300 mg/L spectinomycin, so that the resistant clones could be identified by their green color. The green embryogenic calli were then converted into embryos on a suitable medium in the presence of spectinomycin (150 mg/L). After 2 months on the embryo induction medium, the embryos were transferred to an embryo-germination medium containing spectinomycin (150 mg/L). Interestingly, soybean is naturally resistant to high concentrations (800 mg/L) of streptomycin. Plastid transformation in soybean is a good example for combining a crop-specific plant regeneration protocol with spectinomycin color selection. Another salient feature of the soybean system is the absence of spontaneous spectinomycin-resistant mutants. This may be the case because the mutations that would confer spectinomycin resistance are not compatible with ribosome function. The third salient feature of the soybean system is the absence of wild-type ptDNA copies in the regenerated plants, as in potato. The uniform population of transformed ptDNA copies in the regenerated plants is likely to be due to protracted cultivation on the selective medium prior to plant regeneration. Construction of insect resistant (Dufourmantel et al. 2005) and herbicide resistant (Dufourmantel et al. 2007) transplastomic soybean plants confirmed the utility of plastid transformation in soybean. Soybean is the most important agronomic crop in which reproducible plastid transformation is currently available.

G. Alfalfa: Medicago sativa

Plastid transformation of alfalfa has been accomplished in Shaochen Xing’s laboratory using biolistic delivery of a homologous, aadA-carrying vector to leaves (Wei et al. 2011). The tissue culture system for cv. Longmu 803 used a typical multi-stage medium for embryo induction, multiplication, germination and rooting. Selection was carried out in the presence of 500 mg/L spectinomycin. Because alfalfa is edible and is used as feedstuff to livestock, it is a suitable crop for oral delivery of vaccines and therapeutic proteins.

H. Lettuce: Lactuca sativa

Two groups reported plastid transformation in lettuce at about the same time. Cilia Lelivelt, Jackie Nugent and a group of collaborating researchers from Rijk Zwaan Breeding, B.V., Fijnaart, The Netherlands and The National University of Ireland, Maynooth, reported plastid transformation in cv. Flora. Transformation was carried out with a homologous lettuce vector carrying an aadA gene that was introduced into protoplasts by PEG treatment. Transplastomic clones were identified on a medium containing 500 mg/L spectinomycin.

Kanamoto and colleagues (Kanamoto et al. 2006) described plastid transformation in cv. Cisco after biolistic delivery of a lettuce-specific vector into leaves carrying an aadA gene, and shoot regeneration on spectinomycin-containing medium. The levels of selective spectinomycin concentrations were 10× lower than in tobacco, 50 mg/L. The efficiency of plastid transformation was comparable to tobacco (one transplastomic clone per bombarded sample), and the level of GFP was very high, ∼36% of total soluble cellular protein. In the meantime, the same group (Lim et al. 2011) has extended plastid transformation to a different lettuce cultivar, Romana.

By 2010, the Daniell laboratory developed an efficient transformation and regeneration protocol for cv. Simpson Elite (Ruhlman et al. 2010). Contributing to the success were (1) adoption of native targeting sequences and regulatory sequences, and (2) cultivar-specific optimization of the regeneration medium to produce transplastomic shoots by direct organogenesis. The Daniell group successfully used the lettuce system for the expression of various recombinant proteins (Ruhlman et al. 2007; Boyhan and Daniell 2011; Kanagaraj et al. 2011). The advantage of the system is that lettuce is edible raw, thus it is suitable for oral delivery of therapeutic proteins and vaccines.

I. Cabbage: Brassica oleracea and Other Species in the Brassicacae Family

The plastids in several species in the mustard (Brassicaceae) family have been the targets of plastome engineering. The plastid transformation vectors carried aadA genes and identification of transplastomic clones was based on spectinomycin resistance. Selective concentrations of spectinomycin, in most cases, were lower (10–60 mg/L) than the concentrations used in tobacco (500 mg/L).

Plastid transformation of oilseed rape (Brassica napus) was carried out by bombardment of green cotyledon petioles (Hou et al. 2003) or cotyledons of cv. FY-4; (Cheng et al. 2010) and selected on a medium containing 10 mg/L spectinomycin. The regenerated plants were heteroplastomic, a problem that can be addressed by repeated cycles of plant regeneration, segregating away the wild-type copies in the seed progeny, or choosing alternate insertion sites to avoid interference with flanking genes.

Genetically stable, homoplastomic lines have been described in two other species in the Brassicaceae family. Plastid transformation in cauliflower (Brassica oleracea var. botrytis) by PEG treatment and selection on 20–60 mg/L spectinomycin yielded a single homoplastomic plant (Nugent et al. 2006). Plastid transformation has also been reported in Lesquerella fendleri (Gray) Wats A14581, a species with a desirable seed oil composition and a high capacity for plant regeneration from leaves (Skarjinskaia et al. 2003). Leaf bombardment with aadA vectors and selection on spectinomycin (400 mg/L) yielded fertile, homoplastomic plants. Plastid transformation was relatively inefficient: in 51 bombarded leaf samples, only two transplastomic clones were obtained, possibly due to the use of heterologous vectors. Surprising was the large number (110) of spontaneous mutants in the experiment.

Systematic research in the laboratory of Menq-Jiau Tseng led to the establishment of a reproducible system for plastid transformation in cabbage (Brassica oleracea L. var. capitata L.) (Liu et al. 2007). The protocols have been implemented in cultivars K-Y cross and Summer Summit. Biolistic DNA delivery with homologous, aadA-containing vectors was followed by initial selection on 50 mg/L spectinomycin, followed by cultivation on 200 mg/L spectinomycin. The utility of plastid transformation in cabbage was demonstrated by the expression of insecticidal cry1Ab protein gene in chloroplasts (Liu et al. 2008).

J. Thale Cress: Arabidopsis thaliana

Arabidopsis thaliana is also a member of the mustard family (Brassicaceae). We obtained plastid transformation in Arabidopsis by combining the tobacco leaf transformation protocol with the two-step (callus induction, plant regeneration) Arabidopsis tissue culture and plant regeneration protocols (Sikdar et al. 1998). Because the leaf cells in Arabidopsis are polyploid, we obtained sterile plants. However, the meristematic cells in a shoot apex or cells of a developing embryo are diploid. To maintain the diploid state in our culture, we developed an embryogenic culture system for plastid transformation in Arabidopsis by regulated expression of the BABY BOOM transcription factor (Lutz et al. 2011). This investment has yet to yield fertile transplastomic plants.

K. Sugar Beet: Beta vulgaris

Plastid transformation in sugar beet was reported from Michele Bellucci’s laboratory following biolistic DNA delivery to leaf petioles using a homologous vector, and selection in the presence of 50 mg/L spectinomycin (De Marchis et al. 2009). Interestingly, like soybean, sugar beet is also naturally resistant to high concentrations (1,000 mg/L) of streptomycin. The transplastomic clones appeared after 5 months of selection in the Z025 line. Plant regeneration was obtained only after spectinomycin was removed from the medium. The regenerated plants were heteroplastomic; however, two additional rounds of shoot regeneration in the presence of low (12.5 mg/L) spectinomycin concentrations yielded homoplastomic plants. Overall, in this first experimental series, it took 14 months to obtain transplastomic sugar beet plants. Sugar beet is an important industrial crop of the temperate zone in which chloroplast DNA is not transmitted through pollen, like in most flowering plant species. Plastid localization of transgenes could alleviate concerns about gene flow in the field due to the well documented cross-compatibility of sugar beet with its wild relative sea beet (B. vulgaris ssp. maritima; (De Marchis et al. 2009)).

L. Carrot: Daucus carota

Transplastomic carrot (Daucus carota cv. Half long) was reported from the Daniell laboratory (Kumar et al. 2004b). Transplastomic carrot was obtained after biolistic DNA delivery of a homologous vector carrying an aadA gene, and selection for increasing concentrations (150, 350 and 500 mg/L) of spectinomycin. The transgenic calli had a green colour, attributed to the expression of badh transgene introduced by linkage to the aadA gene (see Sect. III.A).

M. Poplar: Populus alba

Okamura and colleagues (Okumura et al. 2006) reported plastid transformation in poplar after biolistic DNA delivery of a homologous vector carrying an aadA marker gene. The vector DNA was introduced into leaves, and the transplastomic shoots were recovered by selection for spectinomycin resistance (30 mg/L). A significant number of spontaneous plastid-encoded spectinomycin resistant mutants were also obtained. Poplar is a potential biofuel crop, in which plastid transformation may be useful to improve the value of the crop by co-expression of value-added products.

N. Cotton: Gossypium hirsutum

Plastid transformation in cotton was reported from the Daniell laboratory using cv. Coker 310FR (Kumar et al. 2004a). Because spectinomycin was reportedly toxic, after biolistic DNA delivery, the selection of transplastomic clones was carried out on kanamycin. The selective concentration of kanamycin was initially 50 mg/L and then increased to 100 mg/L in subsequent cycles. The Double Gene/Single Selection vector carried the aphA-6 and nptII genes. Because both genes confer resistance to kanamycin and expression of neither of the genes alone has been tested, the rationale behind the approach remains unclear. Although of significant potential economic interest, plastid transformation in cotton has not yet been duplicated nor have transplastomic seeds been distributed for analyses.

O. Cereals: Rice (Oryza sativa) and Wheat (Triticum aestivum)

Cereal plastids are naturally resistant to spectinomycin due to having the 16S rRNA nucleotide substitution that confers spectinomycin resistance to sensitive ribosomes (Fromm et al. 1987). Therefore, we attempted selection for streptomycin resistance with homologous rice vectors carrying an aadA gene (Khan and Maliga 1999). Biolistic delivery of the transformation vector into cultured embryogenic cells was followed by selection on streptomycin-containing plant regeneration medium. Because AAD, the aadA gene product, was fused with GFP, chloroplast localization of the fusion protein could be detected by fluorescence microscopy. However, in the absence of repeated cycles of plant regeneration, we did not obtain homoplastomic plants. Lee and colleagues (Lee et al. 2006) duplicated the experiment and carried it a step further by demonstrating that the transformed plastids can be transmitted into the next generation. However, they could not find a solution to the problem of obtaining homoplastomic plants from the cultured rice cells.

Chloroplast transformation in wheat (Triticum aestivum L.) was reported recently (Cui et al. 2011). The transformation vector was introduced into immature scutella and inflorescences by the biolistic process, and transplastomic clones were selected by resistance to 30, 40 and 50 mg/L of G418 in the first, second and third selection cycles, respectively. Bombardment of ∼2,500 scutella and ∼600 immature inflorescence sections yielded one homoplastomic and two heteroplastomic plants, a relatively low frequency. Two facts cast doubt on the validity of the claims. (1) Transformation with the vector, as described, results in the deletion of the atpB coding region N-terminus. Deletion of atpB would results in pigment deficiency in other plants. The transplastomic wheat plants were reported to have a green, wild type phenotype. (2) Probing of total plant cellular DNA with the rbcL-atpB targeting region was reported to detect a 2.5-kb BamHI fragment. In the wild type wheat plastid genome (AB042240), the rbcL-atpB region is contained in a 9.5-kb BamHI fragment. Because the artificial BamHI cloning sites from the transformation vector are not incorporated in the transplastomic wheat ptDNA, the 2.5-kb signal in Fig. 4 suggests probing plasmid, rather than total plant cellular DNA. If the reported data are true, we shall soon see confirmation of these findings from multiple laboratories.

VII. Perspectives

Plastid transgene expression offers many advantages, but it is still barely utilized. The obvious reason is that, more than 20 years after its first implementation, the technology is still not available in most crops. What needs to be done to accelerate progress?

Relative uniformity of plastid genomes and the acceptance of heterologous vectors within the Solanaceae led us to believe that we did not necessarily need species-specific vectors. The recent example of efficient potato plastid transformation being dependent on homologous vectors is a wake-up call (Valkov et al. 2011). If significant intraspecific sequence diversity turns out to be the rule, we may need to develop multiple vectors for each species, dependent on the tolerance of its recombination system for sequence variation. This calls for more plastid genome sequencing. Fortunately, next generation sequencing provides the tool to rapidly determine the plastid genome sequence from total plant DNA of the cultivar we intend to transform (Nock et al. 2011) and, based on the sequence, we can decide if construction of line-specific vectors is justified. If construction of homologous vectors is required, we can replace cloning by purchasing synthetic targeting regions.

Spectinomycin selection was useful to recover transplastomic clones in many species. Even if spectinomycin selection is feasible in a crop, we need at least one additional marker for multistep engineering. However, some of the crops, such as the cereals, are naturally resistant to spectinomycin. Just finding the right antibiotic or selectable marker gene may solve the problem of obtaining homoplastomic plants in monocots. That is why testing a wider array of antibiotics, for example G418 to which the neo (nptII) and aph(3′)IIa genes confer resistance (Sect. III.A), and new selectable marker genes may be important for extending plastid transformation to new crops.

Plastid transformation is available for the expression of recombinant proteins in the well-established tobacco system and the newly developed, edible hosts lettuce and alfalfa. Extension of the technology to new crops would significantly enhance its utility. The most desirable agronomic application would be containment of herbicide-resistance transgenes and disease resistance traits in the wind-pollinated cereal crops.